Sand Deposition and Poverty among Farm Households

Farm Productivity Loss due to Flood-Induced Sand Deposition: A Study in Dhemaji, India

Kalyan Das

Working Paper, No 73–12

Published by the South Asian Network for Development and Environmental Economics (SANDEE)PO Box 8975, EPC 1056, Kathmandu, Nepal.Tel: 977-1-5003222 Fax: 977-1-5003299

SANDEE research reports are the output of research projects supported by the SouthAsian Network for Development and Environmental Economics. The reports have beenpeer reviewed and edited. A summary of the findings of SANDEE reports are alsoavailable as SANDEE Policy Briefs.

National Library of Nepal Catalogue Service:

Kalyan Das


(SANDEE Working Papers, ISSN 1893-1891; WP 73–12)

ISBN: 978-9937-596-02-2

Key words: Valuation

Production function

Sand deposits

Soil quality

Floods

India

SANDEE Working Paper No. 73–12


Kalyan DasOKD Institute of Social Change and DevelopmentGuwahati, India

October 2012

South Asian Network for Development and Environmental Economics (SANDEE) PO Box 8975, EPC 1056, Kathmandu, Nepal

SANDEE Working Paper No. 73–12

The South Asian Network for Development and Environmental Economics

The South Asian Network for Development and Environmental Economics (SANDEE) is a regional network that brings together analysts from different countries in South Asia to address environment-development problems. SANDEE’s activities include research support, training, and information dissemination. Please see www.sandeeonline.org for further information about SANDEE.

SANDEE is financially supported by the International Development Research Center (IDRC), The Swedish International Development Cooperation Agency (SIDA), the World Bank and the Norwegian Agency for Development Cooperation (NORAD). The opinions expressed in this paper are the author’s and do not necessarily represent those of SANDEE’s donors.

The Working Paper series is based on research funded by SANDEE and supported with technical assistance from network members, SANDEE staff and advisors.

AdvisorE Somanathan

Technical EditorPranab Mukhopadhyay

English EditorCarmen Wickramagamage

Comments should be sent to Kalyan Das

OKD Institute of Social Change and Development

Guwahati, India

Email: [email protected]

Contents

Abstract

1. Introduction 1

2. Valuing Environmental Damage 2

3. Study Area and Data 3

3.1 Soil Characteristics 3 3.2 Data Collection Strategy 4 3.3 Descriptive Statistics 5

3.3.1 Economic Background of the Families 5 3.3.2 State Support for Coping 6

4. Soil Analysis of Agricultural Plots 6

5. Data Analysis 7

5.1 Estimating a Production Function 7 5.2 Examining intra- and inter-district variation 9 5.3 Estimating Damage Cost 9

6. Conclusions and Policy Implications 10

Acknowledgements 11

References 12

TablesTable 1: Summary Statistics of 1,059 Households 14Table 2: Average Monthly Household Income from Agriculture in Rupees 14Table 3: Yield of Paddy per hectare (in kg) in Self-Owned Plots 15Table 4: Soil Characteristics of the Sampled Paddy Plot 15Table 5: Sand Magnitude (%) in the study Villages 15Table 6: Summary Statistics of Variables Used in Regression 16Table 7: Regression Results of Determinants of Paddy Yield 17Table 8: Estimates of losses per hectare from increased sand intensity 17

List Figures

Figure 1: Map of Dhemaji 18Figure 2: Sand deposited area of Dhemaji district 19Figure 3: Sand deposited villages of Jiadhal Basin Area, Dhemaji district 19Figure 4: Trend of yield of Rice in Dhemaji district and State of Assam (1990-91 to 2009-10) 20Figure 5: Sand as a determinant of paddy yield (all paddy plots) 20Figure 6: Sand as a determinant of paddy yield (paddy plots using HYV seeds) 21Figure 7: Study area-land plots and embankment breaching points 2009 22

Appendix Tables 23

Table A1: Nutrition Availability in Differently Textured Soil 23Table A 2: Correlations Coefficients (only significant values reported) at 95% Confidence Level 23

Annextures

Annexure-I: Household Survey Schedule for Listing (stage- I) 24Annexure-II: Household Schedule for Detailed Survey 25

South Asian Network for Development and Environmental Economics6

Abstract

It is widely believed that the decline in agricultural productivity in

the Dhemaji district of Assam, India, is due to flood-induced sand

depositions in paddy fields. Increased sand content reduces the water

retaining capacity of soils and reduces crop productivity. Soil tests in

346 agricultural plots in the Jiadhal basin of the Dhemaji district show

high concentrations of coarse sand, low pH, low organic carbon and

nutrients. Some 39% of the plots tested had sand concentrations greater

than 70%. An analysis of the effect of sand on productivity suggests

that sand deposits have a significant effect on paddy yields. The annual

damage costs from sand deposition in the study area are estimated to

range from INR 690 to 1845 (US$ 13 to 35) per hectare. This study

underscores the need for further research to identify other factors that

may also be responsible for the decline in paddy productivity in the

region, while proposing interventions to improve paddy productivity.

Key Words: Valuation, Production function, Sand deposits, Soil quality,

Floods, India.

1



1. Introduction

Studies have established soil quality to be an important factor in explaining farm productivity (Jaenicke and Lengnick, 1999; Andrews and Carroll, 2001). Although farm size1, labour and other inputs such as fertilizer and mechanization are important factors to be considered in any explanation of variability in farm productivity, soil quality is indubitably the most significant (Bhalla and Roy, 1988). Soil quality is also understood to play a role in ensuring the quality of groundwater and surface water (Doran and Parkin, 1994)).

In floodplains across the world, farmers have been using floods as a traditional mechanism to upgrade soil quality because they enable silt deposition (D’Souza 2006, Mishra 2001, Dixit 2009). Although floods are known to cause substantial damage through destruction of standing crops, property, human life and livestock, among the positive impacts of floods is the deposition of alluvium (or silt) which replenishes the productivity of soil. However, the deposition of a sandy layer, which can have long term impacts on soil and soil fertility and may even be irreversible2, is a less talked about phenomenon associated with floods. Since the damage caused by the deposition of a sandy layer is not a widespread issue (Reddy, 2003), it receives little space in discussions on floods and policy responses to floods in the state of Assam.

Dhemaji district (see Figure1), once the rice bowl of Assam, is now transformed into a virtual desert due to sand deposition as a result of flooding from the Himalayan tributaries of the Brahmaputra River (Figures 2 & 3). Satellite imagery from 1999 reveals that over 11,247 hectares have been rendered unproductive by sand deposition in the district (National Resource Management Foundation, 2006). Further, between 1992 and 2004-5, the net sown area in the district decreased by about 11% (7,689 hectares) and fallow and uncultivated land increased by 35% (8013 hectares) (GoA, 1996 and 2008).

A Report by ICAR (1980) indicates that the average productivity of paddy in Dhemaji was the same as the state average in 1980-81 (about 1210kg/ ha). However, as show in Figure 4, average paddy productivity in the Dhemaji district is now much lower (1,290kg/ha according to 2009-10 data) relative to the state average (1,765 kg/ha). In our study area, information derived from a near-complete enumeration3 of 1,059 households in 15 sample villages, reveals that nearly 83% of the total paddy acreage (1,510 hectares), faces sand deposition. In addition to the productivity losses, there is concern about possible biodiversity losses. Dhemaji district produces numerous indigenous varieties of paddy that may disappear as paddy yield declines. In the 15 villages covered under this study, we recorded 13 indigenous varieties of paddy.

Sediments transported by rivers are of two types: (a) the bed sediment load of coarser particles, and (b) the wash loads of finer particles (or micro-nutrients). Generally, in high flow conditions, only the finer particles carried by the wash load get deposited in fields (Weibe, 2006). In contrast, floods now carry relatively heavy bed sediments

1 Since the 1960s, studies have established a strong inverse relationship between farm size and land productivity Sen, 1962; Mazumdar, 1965). Bhalla and Roy (1988) argue that once proper account is made of the exogenous land quality variable, the inverse relationship is observed to weaken and, in many cases, even to disappear.

2 This statement is based on our observations of the massive piling up of sand in paddy fields in the Jiadhal basin area of upper Brahmaputra valley during a visit in 2004. In subsequent years, however, the deposition of finer silts over the sandy layers has helped to reclaim land to a certain extent and to obtain some returns on paddy in some villages. The shifting of the river relative to the paddy fields and the distance between the paddy fields and the river bank are also determinants influencing the yield.

3 The Census of India (2001) shows there were altogether 1113 households in the 15 study villages.


(sand)4 in the Dhemaji district. Goswami and Das (2003) have identified shifting cultivation, rampant tree felling for commercial purposes, and the extraction of boulders in the upstream mountain valleys as the reasons for the accelerated rate of sedimentation.5 The damage caused by sand deposition is not as high in other valleys in the north bank of Assam, although they too are inundated by floods.

An important question to ask is what the impact is of these sand deposits and what the resulting costs are. There are three dimensions to the assessment of the cost of damages in such situations: (1) the cost incurred due to depleting housing stocks and standing crops; (2) the cost of damage to land which to a large extent is irreversible; and (3) the cost incurred in the mitigation plans and relief operations of the state and the coping and adaptive strategies adopted by the affected people. However, policy papers, prescriptions and valuations of flood damages generally miss the second dimension, thus failing to arrive at a comprehensive estimation of flood damages.

Studies from other parts of the world have established a link between a decline in paddy productivity and anthropogenic-induced sand deposition (Caldwell et al., 2002). While studies on river flooding and erosion in north-east India address the link between sedimentation and the corresponding rise in river beds, channel shifting and erosion, the damage caused by sand deposition in paddy fields has attracted little attention (Weibe, 2006). The present study addresses this gap. We attempt in our study to value the damages to cultivable land and the loss of productivity due to flood-induced sand deposition.

The rest of the paper is as follows. Section II provides a brief overview of existing methods to value environmental damage. Section III describes the study area and the sampling frame. Section IV discusses soil quality in the flood-affected paddy plots. Section V identifies the factors influencing the productivity of paddy and estimates the damage due to sand deposition using an econometric model. Section VI concludes with a discussion of policy options and adaptation issues.

2. Valuing Environmental Damage

Two approaches are popular in the valuation of environmental damages: the Production Function Approach and the Ricardian Approach. The Production Function Approach takes the environment as an input in the production of a marketable good (Vincent, 2011) and estimates its potential positive or adverse effects (Kumar and Parikh, 1998). However, the traditional production function approach is known to have inherent biases with regard to damage valuation because it overvalues the damage since it omits the diverse adaptations that farmers customarily make in response to changing economic and environmental conditions.

Production function studies often do not consider the introduction of a new crop in response to the changing physical conditions of the land, technological changes, and changes in land use. Critics feel that this bias can be corrected by the Ricardian Approach, which accounts for farmers’ responses to various climatic or environmental conditions for the purpose of profit-maximization. On the other hand, the Ricardian Approach assumes that, under competitive conditions, the value of land or the land rent will be equal to the net yield of the highest and best use of land. In agriculture, land rent tends to be a large fraction of the total costs, which can be estimated with reasonable precision (Mendelsohn et al., 1994). We may capture land value through net revenue (Kumar and Parikh, 1998), through farmers’ own perceptions regarding the value of the land (Maddison et al., 2007), or through the actual sale value of the land.

Researchers have, however, expressed reservations regarding the use of sale price in the Ricardian Approach because it requires the incorporation of several supportive parameters which could influence the value of land (Koundouri and Pashardes, 2003). The availability of water, for one, could revise the value of land upwards despite other adverse environmental conditions (Mendelsohn and Dinar, 2003). Furthermore, in any given year only a very

4 In the study area, we observed massive sand deposition in the villages, where the embankments have been breached in recent times (see Figures 3 & 7). This may open up the debate on the need for effective embankments. However, in recent times, there have been few studies advocating embankments except for a study by Myaux et al. (1997) that argues, in a different context, that river embankments contribute to the health status of children in flood-prone areas.

5 Our study did not attempt to have a measure of these activities due to unavailability of secondary data.

3


small fraction of farmland is likely to be put on the market making the sample small and unrepresentative.6 Keeping these methodological challenges in mind, we now proceed to a description of the study area. As we discuss later, given limited land transactions in our study area, we use the production function approach to estimate damage costs borne by farmers from sand depositions.

3. Study Area and Data

Dhemaji, which is a narrow valley district that is 2,637sq.km in area and approximately 30 km in width located in northeastern Assam, is squeezed by the steep slope of the Arunachal Himalaya from the north and the east, and by the Brahmaputra River in the south (See Figure 1). Its catchments receive the heaviest rainfall (on average 3,500 mm annually) while the 19 tributaries of the Brahmaputra carry heavy loads of sediment (Goswami, 1998).

The district is home to 1,140 villages and a population of 688,000. Floods in the year 2000 affected at least 333,000 people in 810 villages, damaging 11,331 hectares of standing crops. As reported by the state disaster management authority, the flood in 2011 had affected a population of 154,500 and an area of 28,295 hectares in 261 villages of the district.

The Government of Assam spends millions of rupees annually in disaster avoidance, primarily for the construction of embankments, relief operations and mitigation activities. To date, 40 embankments and dykes have been constructed in the district while another six have been sanctioned costing INR 1,016.38 million (DRDA, 2009).

3.1 Soil Characteristics

Although detailed historical records on the condition of the soil in the study area are not available,7 there are indications that not so long ago the soil of this flood-affected area was conducive to high rice productivity.

The soil of the district is mainly alluvial with a mixture of sand and clay in varying proportions. The soil is deep and well drained with a varying texture that ranges from coarse loamy to fine loamy. The district has large expanses of low-lying land with heavy clayish soil and a high percentage of nitrogen, which is conducive to rice cultivation. The new alluvial soil of the study area was subject to recurrent floods carrying annually large loads of silt (ICAR, 1980). The soil is rich in organic content and, consequently, high in nitrogen, which is a typical feature of soil in this area.

The National Bureau of Soil Survey provides information on soils for three areas of Dhemaji: two areas of Bordoloni and one area of Maralgaon (NBSS, 1994). According to available data, the soil texture of the Bordoloni area was sandy with alluvial as parent materials. The land capability is classified as III ws, which means wet cultivable land with rooting zone limitation.8 The texture of soil at other points of the Bordoloni area shows itself to be coarse silty with alluvial as parent material. The area is severely flood-prone with land capability classified as III w. The texture of soil at Maralgaon shows itself on the other hand to be coarse loamy alluvial with land capability level identified as III ws with a soil depth of more than 100 cm. This suggests that the area requires the management of excess water while crops must be selected that are adapted to the wet conditions. The pH value of soil in all these areas found to be in the ideal range of 6.5 to 7.0 (NBSS, 1994). The areas covered in the NBSS samples include some sample villages included in this study (see Figure 7).

The issue of severe flood damage and land degradation in the study area came into focus only in the late 1990s. A study commissioned by the State Land Use Board, Assam, indicates that 14,261 hectares of paddy in 575 villages were damaged by floods in 2003 (NRMF 2006). In the following year, the paddy area that was damaged was approximately 21,388 hectares in 465 villages. The study used remote sensing data which indicated three distinct 6 Expectation of future development could be a reason for holding back the land. In our study area however the reason is that very few

value the sand deposited degraded land, the damage which is considered as irreversible. 7 The reason for this is because government labouratories in the State of Assam test the texture of the soil by the feel and touch method,

which cannot correctly indicate the exact proportion of sand, silt or clay in the soil. In our study, the soil tests were conducted in a private labouratory.

8 The land capability classification of the US Department of Agriculture allocates land suitable for cultivation to classes I to IV; grazing to classes V-VI; forestry to class VII; and wildlife and recreation to class VIII. These classes are divided into sub-classes of erosion risk (e), wetness (w), rooting zone limitation (s), and climatic limitation (c) (Dent and Young, 1981).


clusters of sand deposits in the district by the three tributaries of Brahmaputra: Jiadhal, Gainadi and Simen rivers (see Figure 2).

The shift in the course of the river Jiadhal and the construction of embankments have given the farmers some respite from the destruction caused by numerous floods. In some villages, embankments have prevented the entry of coarse sand carried by the first waves of the flood to the paddy fields although subsequent flood waves, which overflowed the embankments, deposited silt that enriched the degraded fields.

Soil scientists have shown that paddy cultivation is possible on silted land, even where there is a high proportion of sand at the base, if water supply is assured throughout the growing period. In the case of Dhemaji, although the coarse texture of the soil reduces the water- and nutrient-retaining capacity of land, farmers have the advantage of the relatively long monsoon period from June to October to overcome the lack of irrigation. Moreover, rice varieties which have a shorter ripening period enable farmers to get a reasonable return. But the soil texture9 influences transmission and storage of water, the flow of air in the soil, and the soil’s capacity to supply nutrients. Because fine textured soils are in general more fertile than coarse textured soil (i.e., sandy) primarily due to the higher clay content and organic matter, soils of fine and medium texture are commonly in use in rice cultivation (See De Datta, 1981). A higher content of sand, on the other hand, leads to lower productivity of the soil.

3.2 Data Collection Strategy

We selected for our study the area between the twin rivers of Kumotia and Jiadhal in five flood-affected Gram Panchayats of Dhemaji district (see Figure 7). Satellite imagery data indicate this area to be one of the major sand-deposited areas (see Figure 5). We selected 15 villages randomly from the 93 villages in the five Gram Panchayats.

We collected the data in two stages. In Stage I, we administered a one-page questionnaire to almost all the households (1,059 households in all) of the 15 selected villages. We collected information on the possession of agricultural land, agricultural land affected by sand deposition, land leased for cultivations and the terms of the lease, the quality of the land leased, and land purchased and sold in recent times for the purpose of understanding land value in a degraded environment, the total production of paddy, conversion of agricultural land for other economic activities, and out-migration from the households. We estimated the average plot size at 1.4 hectares per household. The data showed sand deposition in 84% of the total paddy growing area. The leasing of land for cultivation purposes and out-migration of family members in search of jobs from the households, which are indicators of distress (Nair and Ramanathan, 2007), have become significant phenomena in the study area (Table 1). We found land transactions to be virtually absent in nine of the 15 sample villages, indicating that people do not value degraded land.10

In Stage II, we randomly selected 10 households from each village for the purpose of obtaining more detailed information on household characteristics, agricultural operations including input levels of mechanization, land improvement measures, access to credit, migration, and extent of state support.

In addition, we collected soil samples from every agricultural plot (346 in all) of the 148 sampled households. We measured the texture of the soil, pH factor, organic carbon, and the nutritional contents of the selected categories of texture in order to assess the quality of the soil for the purpose of relating it to the returns from agricultural operations.

We collected soils up to 15 inches in depth because paddy roots supposedly penetrate at least 15 inches into the soil.11 For each plot, we collected soil samples from a minimum 4 to a maximum 16 grid points depending on the

9 Texture refers to the size of the soil particles. Texture is the relative proportion of sand, silt and clay in each soil. Texture classes are: i) coarse texture sands, loamy sand and sandy loams with less than 18% clay, and more than 65% sand; ii) medium texture sandy loams, loams, sandy clay loams, silt loams with less than 35% clay and less than 65% sand. The sand fractions may be as high as 82% if a minimum of 18% clay is present. For more details, one may refer to the soil texture triangle followed by the US Department of Agriculture which is available in standard textbooks like Ashman and Puri (2002).

10 Possible compensation from the state for degraded land is minimum. The revenue department of the state has provision to allot INR 6000/ per bigha or INR 45,000 per hectare for removal of sand from the paddy plot. Considering the magnitude of sand deposition the amount is considered to be insufficient.

11 Rice roots do not penetrate very deep in rain-fed lowland conditions. Depending on the soil conditions, roots may penetrate between 15 and 25cm in depth in rain-fed areas (Jamson, 2011).

5


size and uniformity of the plots, after making certain subjective assessments. We mixed up soils collected from different grid points of a plot to make the sample representative.

We measured the texture of the soils, the pH factor, and the organic content of all 346 sampled soil plots.12 The pH range of 6.3 to 6.8 is considered ideal for paddy cultivation. The high organic content of the soil reflects the availability of nitrogen in the soil. The texture, pH value and organic content would reflect the status and suitability of the soil for paddy.

3.3 Descriptive Statistics

3.3.1 Economic Background of the Families

The data showed the average land holding to be 1.4 hectare per family although their mean monthly return from agriculture was estimated to be only INR 305. Since little of the output is sold in the market, we multiplied the annual production by the average wholesale market price in order to obtain the annual agricultural income. This return is distressingly low (see Table 2) and production is not adequate to ensure even the 30kg of rice required for a family of five per month.13 The returns from agriculture, however, are relatively better in a few villages that have a locational advantage. Two villages in our study area, Tekjuli and Tingharia, which are located on the upper course of the river, enjoyed slightly better returns.

We found low agricultural incomes to be pushing people to seek certain non-farm jobs. Agricultural income now constitutes less than 30% of the total household income in more than 90% of the sampled households. Although some households had salaried jobs, most relied on income from casual labour, road construction (made available by the Public Works Department of the state and the General Reserve Engineering Force [GREF]), jobs in the transport sector, livestock rearing (mostly piggery), and weaving (in the case of women). The National Rural Employment Guarantee Scheme (NREGA) and remittances from family members working outside the district were an additional source of income. Salaried jobs included school teaching, defense and para-military jobs as well as petty employment in private sector services. Those with regular government jobs enjoyed higher incomes compared to the others.

This decline in agricultural income has triggered large scale migration of residents to distant places such as the plywood and rubber factories of Kerala and private security service jobs in Tamil Nadu. In fact, most out-migrations in the sample were either to destinations outside the district (26%) or to those outside the state (55%). The incidence of out-migration reported by households was more than 20% of the total 1059 listed households. The proportion of out-migrants is significantly higher in this area compared to the rest of the country, where out-migration on average is about 9% (GoI, 2010).14 The average remittance from migrants was INR 2,198 per month in the 33 households in our sample which reported out-migration.

There is some effort to reclaim land for agriculture in the area with roughly one fifth of the households in the 15 villages leasing land for agricultural purposes. Although the returns, as mentioned earlier, were low, the opportunity cost of not planting paddy is high in the poorly diversified economy of the area. The survey data reveals that the yield of paddy in leased plots is 503 kg per hectare. Although this is marginally greater than the reported productivity on one’s own land (at 354 kg/ha), it is not significantly different (see Table 3). Moreover, these numbers are much lower than the average for the state as a whole which stands at approximately 1,635 kg per hectare (GoA, 2010).

12 The pH is a measure to test how acidic or basic a substance is. The pH scale ranges from 0 to 14 where a scale of 7 is neutral, a value less than seven is acidic, and a value greater than 7 is basic alkali.

13 Rice is the staple food in the region and constitutes all three meals of the day. According to the Ministry of Agriculture, Government of India, the per capita consumption of rice stood at 208 grams per capita per day in 2001. The rice requirement for an average family of five therefore comes to approximately 31.2kg of rice per month.

14 The relationship between sand deposit and outmigration is found to be significantly negative in our sample. This paradoxical result can be explained by the fact that there are several reasons for out-migration including push factors such as overall flood damage and pull factors such as higher wages in areas to which the migrants go.


3.3.2 State Support for Coping

The affected villages have been able to secure some livelihood support under the flagship programs of the Government of India such as the Indira Awash Yojona (IAY, the housing program for the poor) and NREGA to name a couple. According to reports, the district administration tries to link the flagship programs to the flood relief support program. Therefore, we found the support received from the state in the form of food grain supplies in the flood-affected villages to be meager. The average estimated value of food grain received by each affected household averaged just INR 1,326 for the entire year, Dihingia being the village to receive the highest, at INR 5,047, because it is located close to the national highways. Considering that the market price of rice is INR 15 per kg and that of lentils is INR 60 per kg, the support received may suffice for rations for just two months for the average Assamese household of five.

The average number of workdays that people were granted under NREGA was just 14 days in the sampled villages, the maximum being 39 days per household in the Samarajan village as opposed to the stipulated provision of 100 days per year. The type of support available from the state is therefore evidently inadequate. Still 74 households out of the 148 in the sample have able to get some work under NREGA. Supplementing their income from NREGA, 29 households were able to push themselves above the poverty line. Leaving aside income from NREGA, 108 households had monthly income of less than the estimated poverty-line income of INR 4,298.15

4. Soil Analysis of Agricultural Plots

As the discussion above shows, the average productivity in the sampled plots was low (at 467 kg/ha). However, there was a significant variation in paddy productivity as indicated by the standard deviation of 232.

Field visits did not find farmers adopting any type of land improvement measures, including application of chemical fertilizers.16 One explanation for this might be poverty; most farmers of the area being poor, they may not have resources other than manpower and bullocks for investment in agriculture. The other explanation might be that there is little motivation on the part of farmers to invest in the poorly textured, flood-ravaged areas. However, some households (approximately 50% of the sample), have started using tractors to till land.

Soils in the study area are poorly textured with a significant presence of sand and silt, which is acidic in nature having a poor organic content (Table 4). A pH level below 6 inhibits the process of nitrification while low organic carbon indicates low nitrogen content in the soil. However, there is significant variation in sand deposits and sand content, suggesting that paddy plots are not uniformly affected (see Table 5).

We can classify the plots into 9 categories according to the soil types (see Table A 1 in Appendix). The first three categories of soil--sandy, loamy sand and sandy loam--comprise more than 60% of the sampled plots indicating low nutrient content. We tested for the availability of Nitrogen, Phosphorus and Potassium for five dominant categories of soil (sandy, loamy sand, sandy loam, loam and silt loam) in 81 sampled plots in the study area. The tests reveal that all five categories of soil, irrespective of the dominance of sand or silt in the soil, show poor nutritional content (see Table A 1).17 The deposition of fine silt during floods has failed to ensure the nutritional content in the soil.

Forty-three percent of the sample plots record zero returns. In fact, in the year 2009, 76 out of the 346 sampled plots were not tilled by the farmers. They have cited floods as the reason for zero returns in 36 of these plots and sand as the reason in the case of the other 112 plots. Both floods and sand destroy the standing crops leading to low returns. Farmers claim that tilling was not possible in sandy plots. While flooding is temporary, sand deposition has a long term impact.

15 Planning Commission of India estimates poverty line of India based on INR. 28.65 per capita per day. Considering the average size of family at 5 the amount comes to INR 4298 a month.

16 Data for Dhemaji district show fertilizer application of 4.2 kg per hectare for the Kharif season during 2009-10 relative to the figure for Assam state as a whole, which is 36.9 kg/ha (Government of Assam, 2010).

17 The availability of Nitrogen in the soil is considered poor if the Nitrogen availability is less than 272kg/ha of land. For Phosphorus it is considered low if it is below 22.5kg/ha while for Potassium the limit is 136kg/ha. The Organic Carbon content in the soil is considered low if it is less than 0.5 (VanLoon et al. 2005).

7


5. Data Analysis

5.1 Estimating a Production Function

It is common to study agricultural output variability with the help of a production function that links output with various inputs such as labour, capital, fertilizer, pesticide, irrigation, etc. It is now common also to include soil quality as an explanatory variable (Ekbom and Sterner, 2008). While soil quality would be an environmental factor, the rest would be physical factors. Our main interest here is to examine the impact of sand deposition on productivity. Our underlying understanding is that productivity is a function of environmental and physical factors and farmer effort.

To examine the impacts of sand deposits on productivity, we measured paddy yield in weight and expressed it as the kilogram return per hectare. We measured sand concentration as a proportion18 while measuring distance in meters (see summary statistics in Table 6). We hypothesized that a plot’s distance from the river would be an important variable in explaining variation in paddy yield. We expected the use of HYV seeds (dummy variable) to be advantageous to paddy production due to two reasons: HYV technology boosts productivity while having a short gestation period, which helps where humidity retention by the soil is low. We also expected paddy yield to be negatively associated with soil characteristics, particularly the deviation in soil pH factor from the normal range (See Table A 2 in Appendix).

The usual factors of production such as labour and fertilizer are likely to be endogenously determined and would give biased coefficients (Vincent, 2011). Endogeneity emerges from the fact that farmers in sand deposited plots may either put extra labour to clear the sand and do the ploughing or use less labour because they think it is not worth the effort. Thus, labour effort is correlated with sand deposits.19 We note further that, except manpower, other inputs in agriculture are not used in the study area. The uncertainty introduced by floods may compel farmers to refrain from using inputs such as capital and fertilizers.

The impact of different inputs on paddy output is modeled as

Y = f (Sand, Ph, Distance, HYV, Plotsize)

Where

Y = Paddy output per hectare in kilograms

Sand = Proportion of sand in the paddy plots

Ph = Deviation of pH factor from the ideal value of 7

Distance = Distance of paddy plot from the river in meters

HYV = Adoption of HYV seeds (dummy variable)

Plotsize = Size of paddy plots in hectare

We attempt three different econometric specifications to test for the validity of the impact:

a) Linear version

Yi = b0 + b1 Xi + error

b) Log-Log function in Cobb Douglas form

Ln (Yi) = b0 + b1 Ln (Xi) + b2 Zi + error

c) Tobit Model: In the 346 sampled plots, we recorded 148 plots with zero output in 2009. In order to ensure the validity of our OLS results, we also tested a Tobit model with left censoring at 0. The standard Tobit model assumes that

18 Soil samples collected from paddy plots were tested in labouratory to determine the sand proportion as well as other soil characteristics.

19 Our field data show significant negative correlation (-0.15 at 1% level of significance) between labour hours used in paddy plots and sand deposit level in the plots.


Let,

Yi * = b0 + b1Xi + error

y = 0 if y * < 0

y = y* if y * > 0

We present the regression results in Table 7. Models 1-4 refer to OLS estimations, while models 5-8 are Tobit estimations. The estimates presented here are tested for significance with robust standard errors in order to avoid problems of heteroskedasticity. For the Tobit results, we present only the marginal effects for comparison with the OLS results.

We present results from multiple models in order to assess the robustness of our estimates to changes in specifications. Models 1 and 5, in which sand is regressed on paddy yields, establish the strong negative relationship between the presence of sand and productivity. Models 2 and 6 estimate the impact of sand on productivity but add additional previously-discussed variables that may independently affect productivity. These include plot size, soil quality, distance from river (or flood potential) and use of HYV seeds.

The linear and log-linear models in Table 7 show that sand has a significant negative impact on paddy yield (paddyhectare). In the linear OLS models (1-2) and Tobit models (5-6), we find that for every unit increase in sand concentration, production in paddy per hectare decreases by 2.36 kg to 6.39 kgs. While the signs of coefficients of the Tobit models (5-6) remain the same as in the OLS, the co-efficient for sand is higher in the Tobit estimations relative to the OLS estimations.

As previously discussed, labour is likely to be an endogenous variable. However, to better understand the effect of labour, we ran two additional regressions, (model 3 and 7 in table 7) with labour as an input (log of total labour hours used in paddy plots). The regression results corroborate the possibility of endogeneity. When labour is included in the production function, the coefficients for sand (-1.94 in model 3 and -2.21in model 7) are smaller relative to the rest of the models when labour is not included. Thus, sand appears to have a direct effect on productivity and an indirect effect via labour. Further, the coefficients associated with labour are insignificant. Since, our primary interest is the overall effect of sand, we use the coefficient of sand without labour to value damage costs.

Models 4 and 8 regress the log of paddy per hectare on independent variables. Here, the coefficient of (log) sand concentration is not significant. This indicates that the elasticity of productivity with respect to sand concentration is not different from zero. This is possibly an indicator of such high sand concentrations that a relatively small change in sand concentration is not good enough to lead any change in output.

The hypothesized effects of different independent variables on paddy yield are as expected. Distance between the plot and river (Distriver) showed a positive correlation with the paddy yield. The further the plot was from the river side, the higher the productivity. It is possible that plots nearer to the river get repeated deposition of sand while those further away have better chances of increasing organic concentration over time and recover soil productivity. These may also be closer to human habitations and so people spend more time in the soil recovery processes.

All the regression models show that the size of the plot in hectares (plotsize) has a consistent negative effect, implying that a large plot size leads to lower yield. As discussed earlier, farmers in the study area use largely family labour. Hence, in cases of larger plots, family labour may not be adequate for intensive tillage of the entire land. Our results also show that use of HYV seeds (HYVdummy) contributes to higher yield.

We undertook some non-parametric analysis (i.e., Lowess smoothing), to further examine the relationship between sand and the decline in paddy productivity.20 This analysis, shown in Figure 5, suggests that this relationship is not very strong. We were able to arrive at a somewhat clearer picture of the impact of sand on yield when we smooth the data by taking into account only the 71 plots that had adopted the HYV seeds of paddy (see Figure 6). Some households are now switching to HYV variety of paddy, which has a short gestation period (of 80-110 days) as part

20 LOWESS is a data analysis technique for producing a “smooth” set of values from either a noisy time series or from a scatter plot with a ‘noisy’ relationship between the 2 variables.

9


of an adaptation strategy. This may be because traditional varieties of paddy require longer gestation periods (of 120-150 days), which sandy soil cannot sustain once the monsoon rain withdraws.

5.2 Examining intra- and inter-district variation

While the regression analysis in most of the specifications shows sand as significantly impacting on productivity, the variation in output due to sand is insufficient to explain the intra and inter-district variation in output.21 It is noteworthy that the average yield of rice in the state is 1,433 kg per hectare (which is the average during the two decades of 1990-91 to 2009-10) (GoA, Statistical Handbooks). In comparison, the average yield in the Dhemaji district is 1,165 kg rice.22 In the sand-deposited study villages of Dhemaji district, however, the average yield is 315 kg of rice per hectare.23 In contrast, the flood ravaged Dhubri district in lower Assam has an average yield of 1,304 kg of rice per hectare.24

Time series data of yield for Dhemaji district and for Assam state as a whole show that the gap between the average yield of rice between the state and Dhemaji district has increased over the years (Figure 4). Average productivity of rice during the 10 years period 1990-91 to 1999-2000 was 1,157 kg/ha in Dhemaji district and 1,174 kg/ha in the state as a whole. The gap has increased considerably during 2000-01 to 2009-10 showing average productivity of rice at 1,334 kg/ha for Dhemaji and 1,532 kg/ha for Assam as a whole.

Overall it can be concluded that sand has a significant effect on yield in Dhemaji district. However, our data does not allow us to rule out or examine other factors that may be contributing to the difference in productivity between the study villages and other areas of the state. Local people, however, consider sand as the main determinant of the poor condition of the soil. As previously noted, 50% of the farmers in our sample indicated that sand deposits as the main reason for low yields in their farm plots.

5.3 Estimating Damage Cost

We now proceed to estimate the cost of sand deposition. As mentioned earlier, historical data on soil characteristics in the study area is not available. However, ICAR (1980) provides data on the soil texture of the agro-ecological zone of the middle plain of the north bank of Brahmaputra River. According to this data, the coarse sand composition of soil in the Rangapara (located in same agro-ecological zone of foothill Himalayas as Dhemaji and in the district that neighbors Dhemaji) shows 15.7% sand in the loamy textured soil. Thus, for valuing sand damages, we consider that the normal proportion of sand in soils to be 15.7%.

In our study sample plots, average sand proportions were 54.2% of the soil. Further, soil test shows that 39% paddy plots have more than 70% sand concentration. Regression models 2 and 5 (Table 7) show productivity loss for a one unit increase in sand concentration to be in the range 2.36 to 6.39 kg. If the soils were returned to their normal proportion of sand (15.7%), this would mean a decrease in soil concentration by 38.5%.

In order to obtain damage costs, we predict paddy yields (based on our regression estimation) for sand equal to 15.7% and 54.2% of soil, holding all other variables at their sample mean values. Table 8 shows the predicted paddy yields for these levels of sand for regression models 2 and 5. The decline in paddy as a result of sand deposits is given by the difference between yields. Given the two models, the decline in yield as a result of sand deposits is calculated to fall within the range of 92 kg to 246 kg of paddy per hectare per year.

21 We have tested this considering the average sand proportion in the paddy field of affected villages and average proportion of sand in non-affected villages and putting the sand figures in the regression models. The difference in yield is not much compared to what the figure 4 shows. See section 5.1 for more on this.

22 Although average paddy productivity is low in the Dhemaji district, data for recent years indicate an upward trend (see Figure 4). This may be attributable to the recent rise in paddy cultivation areas (from 19,005 ha in 2004-05 to 40,869 ha in 2008-09) using HYV seeds. Dhemaji district was created in 1987 and we did not get yield data prior to the year 1990. GoA, statistical handbooks of specified years for detail.

23 Paddy production is converted to rice. Field data was collected on the basis of paddy production per hectare. Processing of 1,000 kg paddy yield about 675 kgs rice.

24 In the flood-prone district of Dhubri, many farmers have adopted a risk-averse strategy by an appropriate combination of crops and seasons taking advantage of shallow tube well irrigation. Consequently, there has been an upward trend in paddy yield in the district in recent times (Mandal, 2010). The District however is not affected by sand deposition during floods.


Considering the average price of paddy at INR 750 per quintal at the wholesale market, paddy loss is estimated at INR 690 per hectare (estimate I) and INR 1,845 per hectare (estimate II) (see Table 8). Based on available evidence, we assume that approximately 11,331 hectares in the district have been affected by sand (NRMF, 2006). Thus, annual damage costs for the entire district range from an estimated at INR 7.82 million (US$ 146,138) to INR 20.91 million (US$ 390,760).25

Until the late 1990s, as reported during the field visits farmers in the flood affected area, Dhemaji district produced on average 2,400 to 3,000 kg of paddy per hectare. Considering the present return of 467 kg of paddy per hectare in our sample households, this suggests that yield has declined by 1,933 to 2,533 kg of paddy per hectare in the current period relative to the late 1990s. In monetary terms, this falls between INR 14,498 (US$ 271) (Table 8) to INR 18,998 (US $355) per hectare. Thus, if we compare current average yields to yields based on recall data from our field survey in the 1990s, the loss to farmers in the district is significantly higher than if we use our regression coefficients. However, our estimations based on regression results show the effect of sand alone, other things being held constant, while model III estimations are based on average yields. Thus, while the recall data suggests significantly larger losses, we need to be cautious. These estimates could be subject to recall bias. Further, with model II estimations we are comparing differences in average productivity and factors other than sand may also be responsible for these productivity changes.

GoA (2009) estimated the flood damage in the Dhemaji district to be INR 14,590 per hectare during the year 2007-08. In another study, Weibe (2006) estimates flood damage at INR 6,521 per hectare during the 1999-2005 period. While our valuation numbers focus on sand deposit damages alone and not flood damages, it is still useful to see if they are in line with flood damage estimates. We find that it is appropriate that our estimates of sand deposit damage are smaller than the overall flood damage estimates given above.

6. Conclusions and Policy Implications

The results of the study reflect overall the crisis in livelihood caused by sand induced land degradation in the agrarian economy. It is evident that the near irreversible damage to the paddy fields and poor agricultural outcomes from the partially degraded lands is forcing people to look for other income and livelihood alternatives.

Our regression models show productivity losses range between 92 and 246 kg of paddy per hectare with an average presence of 54% sand in the soil samples. Considering that one kg of paddy costs INR 7.5 and approximately 11,331 hectares in the district have been affected by sand, total annual losses are estimated to lie between INR 8 to 21 million (approximately US$ 150-400,000) in the district of Dhemaji.

Our survey indicates that some households in the District have shifted from paddy to mulberry plantations and pisciculture as an adaptation strategy. However, pisciculture is generally not profitable with acidic soils. Thus, farmers may need to apply significant amounts of lime in order to culture fish and it is not clear how remunerative this practice may be.

Conditions in the Dhemaji area put into stark relief the need to strengthen the social security of the poor. More effective implementation of rural development and employment programs by the government might be one way to ameliorate the rural distress in this region. The state expenditure on the construction and reinforcement of embankments, which has been its main strategy to control floods, has not been as beneficial because of frequent breaches in embankments. Any attempts to improve water management would be helpful to farmers. Sandy soil cannot retain water and assurance of water would improve yields.

Since sand deposition is only a partial explanation for the decline in paddy productivity in the region, our study points to the need for further research to identify other factors that are also responsible for this decline. In addition, it would be extremely useful to understand how better management of upstream areas may result in lower sand deposits in paddy fields.

25 The exchange rate used is 1 US$ to INR 53.5.

11


Acknowledgements

This study received financial support from the South Asian Network for Development and Environmental Economics (SANDEE), Kathmandu. I am grateful to E. Somanathan for research advice and Jean Marie Baland, Pranab Mukhopadhyay, Mani Nepal, and Priya Shyamsundar for valuable feedback. D. N. Chakrovarty was most helpful in clarifying the issues relating to soil samples. I am grateful to the Assam Science Technology and Environmental Council for sharing maps and NBSS data. Special thanks are due to Monjit Barthakur for preparing the GIS map and to Dadul Chutia and Mrinal Baruah for excellent field support. I wish to record my appreciation of the logistical support provided by Anuradha Kafle and Krisha Shrestha of the SANDEE Secretariat.


ReferencesAndrews, SS; Carroll, CR (2001) ‘Designing a soil quality assessment tool for sustainable agro ecosystem

management’. Ecological Applications 11(6): 1573-1585

Ashman, M; Puri, G (2002) Essential Soil Science: A Clear and Concise Introduction to Soil Science. Wiley Blackwell: Oxford

Bhalla, SS; Roy, P (1988) ‘Mis-specification in farm productivity analysis: the role of land quality’. Oxford Economic Papers 40 (1): 55-73

Caldwell, JS; Sukchan, S; Ogura, C (2002) ‘Challenges for farmer- researchers partnerships for sandy soil in North East Frontier of Thailand’. mimeo. Available at: www.fao.org/docrep/ [Accessed on 30th April, 2009]

De Datta, SK (1981) Principles and Production of Rice Production. New York: John Willy and Sons

Dent, D; Young, A (1981) Soil Survey and Land Evaluation. London: George Allen & Unwin

Dixit, A (2009) ‘Kosi embankment breach in Nepal: need for a paradigm shift in responding to flood’. Economic and Political Weekly 44 (6): 70-78

Doran, JW; Parkin, TB (1994) ‘Defining and assessing soil quality.’ In Doran, JW; Coleman, DC; Bezdicek, DF; Stewart, BA (eds.) Defining Soil Quality for Sustainable Development. Madison: Soil Sciences Society of America Special Publication No. 35, pp. 3-21

DRDA (2009) ‘RTI response from office of project director, District Rural Development Agency’. Dhemaji, Dated 12th August, 2009, under Right to Information Act, 2005, Government of India

D’Souza, R (2006) Drowned and Dammed--Colonial Capitalism and Flood Control in Eastern India. New Delhi: Oxford University Press

Ekbom, A; Sterner, T (2008) ‘Production function analysis of soil properties and soil conservation investments in tropical agriculture’. Discussion Paper Series 08-20, June, Environment for Development, Kenya

Goswami, DC (1998) ‘Fluvial regime and flood hydrology of the Brahmaputra River, Assam’. In Kale, VS (ed.) Flood Studies in India. Geological Society of India, Vol. 41, Bangalore, pp 53-76

Goswami, DC; Das, PJ (2003) ‘The Brahmaputra River’. Ecologist Asia Vol. 11 (1):9-14

Government of Assam (1996) Statistical Handbook of Assam Directorate of Economics and Statistics, Guwahati

Government of Assam (2008) Statistical Handbook of Assam. Directorate of Economics and Statistics, Guwahati

Government of Assam (2009) ‘Report on damage of flood released by the Dhemaji district administration for the year 2007-08; website of Dhemaji district administration, dhemaji.nic.in- extracted on 25/6/2009)

Government of Assam (2010) Statistical Handbook of Assam. Directorate of Economics and Statistics, Guwahati

Government of India (2010) Migration in India- 2007-08. Ministry of Statistics and Programme Implementation, June, New Delhi

ICAR (1980) Report of ICAR Research Review Committee for Assam Agriculture University, National Agriculture Research Project, Indian Council of Agriculture Research, New Delhi

Jaenicke, EC; Lengnick, LL (1999) ‘A soil quality index and institutional relationship to efficiency and productivity growth measure: two decompositions’. American Journal of Agriculture Economics 81(4): 881-893

Jamson, BK (2011) ‘Rice root growth function in rain-fed lowlands- implications for rice in uplands’. Drought Frontier Project, International Rice Research Institute, Manila, mimeo

Koundouri, P; Pashardes, P (2003) ‘Hedonic price analysis and selectivity bias’. Environmental and Resource Economics 26(1): 45-56

Kumar, KSK; Parikh, J (1998) ‘Climate change impacts on Indian agriculture: the Ricardian approach’. in A. Dinar et al. (eds.) Measuring the Impacts of Climate Change on Indian Agriculture. Washington, D.C.: World Bank, 140-184

Maddison, D; Manley, M; Kurukulasuria, P (2007) ‘The impact of climate change on African agriculture’. World Bank Policy Research Working Paper, No. 4306, World Bank, Washington, DC

Mandal, R (2010) ‘Cropping pattern and risks management in the flood plain of Assam’. Economic and Political Weekly 45 (33): 78-81

13


Mazumdar, D (1965) ‘Size of farm and productivity: a problem of Indian peasant agriculture’. Economica 32 (126): 161-173

Mendelshon, R; Nordhaus, WD; Daigee Shaw (1994) ‘The impact of global warming on agriculture: a Ricardian analysis’. The American Economic Review 84 (4): 753-771

Mendelsohn, R; Diner, A (2003) ‘Climate, Water and Agriculture.’ Land Economics 79 (3): 328-341

Mendelsohn, R; Dinar, A (2003) ‘Climate, water and agriculture’. Land Economics 79 (3): 328-341

Mishra, DK (2001) ‘Living with floods: people’s perspectives’. Economic and Political Weekly 36 (29): 2756-2761

Myaux, JA; Ali, M; Chakraborty, J; de Francisco, A (1997) ‘Flood control embankments contribute to the improvement of the health status of children in rural Bangladesh’. Bulletin of World Health Organization 75 (6): 533-539

Nair, KN; Ramakumar, R (2007) ‘Agrarian distress and rural livelihoods - a study in Upputhara Panchayat, Idukki District, Kerala’. Working Paper No. 392, Centre for Development Studies, Trivandrum

NBSS (1994) Soils of Assam for Land Use Planning. National Bureau of Soil Survey & Land Use Planning, Nagpur

NRMF (2006) ‘Study of siltation of rivers, its nature, extent and magnitude of problems and their remedial measures in upper catchment of Brahmaputra River in Assam’. Unpublished Report, National Resource Management Foundation, Guwahati

Reddy, VR (2003) ‘Land degradation in India: extent, costs and determinants’. Economic and Political Weekly 38 (44): 4700-4713

Sen, AK (1962) ‘An aspect of Indian agriculture’. The Economic Weekly 14 (4-6): 243-246

Umamaheswari, L; Hattab, KO; Nasurudeen, P; Selvaraj, P (2011) ‘Should shrimp Farmers pay Paddy Farmers? The Challenges of Examining Salinisation Externalities in South India.’ In Haque, AKM; Murthy, MN; Shyamsunder, P (eds) Environmental Valuation in South Asia. Delhi: Cambridge University Press.

VanLoon, GW; Patil, SG; Hugar, LB (2005) Agricultural Sustainability: Strategies for Assessment. New Delhi: Sage

Vincent, JR (2011) ‘Valuing the Environment as a Production Input.’ In Haque, AKM; Murthy, MN; Shyamsunder, P (eds) Environmental Valuation in South Asia. Delhi: Cambridge University Press

Weibe, H (2006) ‘River flooding and erosion in North East India: exploratory consideration of key issues’. mimeo, North West Hydraulic Consultants, Alberta, Canada


Tables

Table 1: Summary Statistics of 1059 Households (15 villages) in Stage I (Primary Survey)

Villages HH Listed Total Agri. Land in the HH, in

hectare

Average Size of Possession in

hectare

% of Agri Land with Sand Deposition

HH Leasing Land for

Cultivation

OutMigration from

the HH

No 1. Tekjuli 42 38.73 0.92 44.6 9 0

Tingharia 71 171.30 2.41 54.5 19 16

Dhiri Chapori 80 234.72 2.92 94.7 15 6

Samarajan 71 82.64 1.16 95.4 20 17

Panitola 29 28.41 0.98 99.5 8 15

Garubandha 94 95.02 1.01 99.1 26 23

Nepalikhuti 20 34.40 1.72 100.0 3 5

Kechukhana 42 89.31 2.13 99.1 10 14

Ratua Bali 87 57.95 0.67 97.9 6 16

Dihingia 153 151.15 0.99 80.2 18 19

Maralchuk 42 59.08 1.41 69.5 3 5

Gohaingaon 73 125.78 1.72 58.2 10 28

Pehiyati 89 91.46 1.03 92.5 13 20

Bhakatgaon 83 98.66 1.19 97.9 11 17

Bahkathika 83 150.95 1.82 78.6 9 12

Average 100.63 1.42 84.1 12 14

Source: Primary Survey (2009-10)

Table 2: Average Monthly Household Income from Agriculture in Rupees (15 villages in Stage II -Sample Survey)

Villages Mean Max Min SD N

No 1. Tekjuli 907.5 2,187.5 400 655.4 10

Tingharia 598.1 , 187.5 687.8 10

Dhiri Chapori 59.4 375 0 120.1 10

Samarajan 153.8 337.5 0 119.6 10

Panitola 714.7 2,062.5 162.5 616.4 10

Garubandha 500 781.3 62.5 223.9 9

Nepalikhuti 17.9 125 0 47.2 7

Ratuabali 128.5 512.5 0 161.9 10

Kechukhana 338.1 606.3 0 185.9 10

Dihingia 150.6 375 0 141.6 10

Maralchuk 97.5 375 0 124.3 10

Gohaingaon 76.3 250 0 90.4 10

Pehiyati 97.4 468.8 0 165.2 12

Bhakatgaon 236.3 500 0 203.8 10

Bahkathika 490.7 875 62.5 281.9 10

All Villages 304.5 2,500 0 148

Source: Primary Survey (2010)

15


Table 3: Yield of Paddy per hectare (in kg) in Self-Owned Plots

Villages Mean Max Min SD N –plots

No 1. Tekjuli 741 1,498 41 452 16

Tingharia 722 1,486 0 510 16

Dhiri Chapori 130 2213 0 537 17

Samarajan 424 1,797 0 492 16

Panitola 1087 2,246 0 544 24

Garubandha 1027 2,582 0 842 17

Nepalikhuti 0 0 0 0 10

Ratuabali 120 743 0 229 18

Kechukhana 321 1,977 0 532 25

Dihingia 179 741 0 258 22

Maralchuk 148 741 0 219 17

Gohaingaon 89 738 0 188 20

Pehiyati 255 1,844 0 552 15

Bhakatgaon 365 2213 0 642 15

Bahkathika 507 1,485 0 475 22

All Villages 425 2,582 0 576 270


Table 4: Soil Characteristics of the Sampled Paddy Plot (N= 346 sampled plots)

Mean Max Min

Ph scale 5.7 7.6 0.25

Organic Carbon % 0.54 2.5 0

Sand in % 54.2 98.5 0

Silt in % 35.7 84.9 0.5

Clay in % 10.1 52.1 0.25


Table 5: Sand Magnitude (%) in the study Villages (N= 346 sampled plots)

Villages Sand Average in % Max Min N-plots

No 1. Tekjuli 66.8 89.0 9 22

Tingharia 46.7 86.5 4.5 18

Dhiri Chapori 74.7 98.9 25 23

Samarajan 57.5 88.5 2.5 25

Panitola 49.8 85 13.5 28

Garubandha 59.6 95.5 1.5 26

Nepalikhuti 81.4 91.9 50.5 12

Ratuabali 60.0 98.8 12 24

Kechukhana 67.6 95 12.5 31

Dihingia 54.6 93 6.5 27

Maralchuk 51.0 96.5 0 18

Gohaingaon 40.4 95 1 22

Pehiyati 62.7 90.5 5.5 21

Bhakatgaon 45.5 92 10 22

Bahkathika 4.7 46 0 25



* Defined as days with temperature deviation more than 5°C above the normal.

Table 6: Summary Statistics of Variables Used in Regression (N=346)

Variables Mean Std. Dev. Min Max

Paddy Yield per Hectare 467 586 0 2995

Sand Proportion in Paddy Plots 54.2 29.9 0 98.5

Deviation of pH from Normal Range of 7 -1.25 0.82 -6.75 0.61

Distance of Paddy Plots from River in mt 816 624.5 0 2525.3

HYV Dummy 0.21 0.40 0 1

Plot Size in Hectare 0.67 0.53 0.13 4.12

Labour hour used 306.7 298.0 0 1708

Log of Paddy yield per Hectare 2.80 0.37 0 3.47

Log of Sand Proportion 1.66 0.36 0 1.99

Log of distance of paddy plots from river 2.75 0.45 0 3.40

Log of plot size in hectares -0.28 0.30 -0.87 0.61

17


Table 7: Regression Results of Determinants of Paddy Yield

Variables OLS Tobit

Model 1 Model 2 Model 3 Model 4 Model5 Model6 Model 7 Model 8

Paddyhectare

Paddyhectare

Paddyhectare

Paddy hectare_ln

Paddyhectare

Paddyhectare

Paddy hectare

Paddy hectare_ln

Sand Proportion in Paddy Plots

-2.64**(1.05)

-2.36** (0.95)

-1.94*(1.18)

-5.14***(1.71)

-6.39***(1.50)

-2.21(1.56)

Deviation of pH from Normal Range of 7

92.90** (37.41)

41.30(45.34)

1.19*** (0.20)

94.53*(57.65)

1.83*** (0.39)

Distance of Paddy Plots from River in mt

0.29*** (0.05)

0.28 ***(0.05)

0.54***(0.07)

0.38***(0.07)

HYV dummy use of hyv seeds

420.89***(67.60)

312.12***(76.02)

2.56***(0.36)

720.59***(93.64)

414.21***(92.18)

3.86***(0.48)

Plot Size in Hectare -152.87***(50.77)

-207.28***(75.70)

-296.83***(101.37)

-382.07***(137.98)

Log of labour hour used

45.99(83.42)

164.32(170.00)

Log of Sand Proportion

-0.092 (0.12)

-0.25(0.22)

Log of distance of paddy plots from river

0.69***(0.12)

1.70***(0.28)

Log of plot size in hectares

-0.20 (0.21)

-0.35(0.36)

Constant 609.85***(64.69)

257.28*** (94.43) 330.84

(205.63)*

-2.48***(0.85)

451.93***(105.23)

135.24(142.3) -131.63

(396.66)

-10.98***(1.93)

Sigma 907.67***(53.10))

744.45***(51.28)

644.27(45.72)

4.14***(0.20)

Observations 346 346346

346 346 346346

346

R-squared 0.02 0.290.21

0.35 0.00 0.040.02

0.10

Robust Standard Errors; *** p<0.01, ** p<0.05, * p<0.1R-squared for Tobit models (5-7) should be read as pseudo R-squared in Table 7. Figures in brackets are standard errors.

Table 8: Estimates of losses per hectare from increased sand intensity

Estimate Per hectare paddy production

(Sand average 15.7 % in paddy plots)

Per hectare paddy production

(Sand average 54.2 % in paddy plots)

Loss of paddy per hectare in kg

Loss per hectare in INR

Model I 680 588 92 690

Model II 428 182 246 1845

Model III. 2400# 467* 1933 14498

Note: Estimates I is based on OLS model 2 and II is based on Tobit model 6 (see Table 7 and Umamaheshawari (2011)). Estimate III is based on recall information revealed by farmers during field visit. # Recall based reports of yield in the late 1990s prior to land degradation, * actual yield recorded from the field survey.


Figures

Figure 1: Map of Dhemaji

19


IRS-1D-LIII, 14 TH FEB,2002IRS-1D,L-III 14 February 2002

Figure 2: Sand deposited area of Dhemaji district (courtesy, ASTEC, Guwahati)

Figure 3: Sand deposited villages of Jiadhal Basin Area, Dhemaji district (courtesy, ASTEC, Guwahati)

IRS-1D,L-III14 FEB,05

IRS-1D,L-III 14 February 2005

Sand Deposited Villages of Jidhol River


3

Figure 4: Trend of yield of Rice in Dhemaji district and State of Assam (1990-91 to 2009-10)

Figure 5: Sand as a determinant of paddy yield (all paddy plots)

010

0020

0030

00pa

ddyh

ecta

re

0 20 40 60 80 100sand

bandwidth = .8

Lowess smoother

3



010

0020

0030

00pa

ddyh

ecta

re

0 20 40 60 80 100sand

bandwidth = .8

Lowess smoother



21


4

Figure 6: Sand as a determinant of paddy yield (paddy plots using HYV seeds)

050

010

0015

0020

0025

00pa

ddyh

ecta

re

0 20 40 60 80 100sand

bandwidth = .8

Lowess smoother

Figure 6: Sand as a determinant of paddy yield (paddy plots using HYV seeds)


Figure 7: Study area- land plots and embankment breaching points 2009

Figure 7: Study area-land plots and embankment breaching points 2009

23


Table A 2: Correlations Coefficients (only significant values reported) at 95% Confidence Level

Variables PaddyYield

Sand Silt OC Clay pH dev7 distriver HYV dummay

plotsize

Paddy yield per hectare 1.00

Sand proportion in paddy plots

-0.14 1.00

Silt proportion in paddy plots

-0.81 1.00

Organic carbon -0.48 0.54 1.00

Clay -0.81 1.00 0.54 1.00

Deviation of pH from Normal Range of 7

-0.26 -0.29 0.26 0.32 0.26 1.00

Distance of Paddy Plots from River in mt

0.40 0.26 1.00

HYV Dummy-use of HYV seeds

0.31 0.13 -0.15 -0.14 -0.15 0.12 1.00

Plot Size in hectare -0.19 -0.12 -0.14 1.00

Appendix Tables

Table A1: Nutrition Availability in Differently Textured Soil

Soil Type Number of Plots

NPK Tested Plot Average N kg/ha Average P kg/ha Average K kg/ha

Sandy 44 11 138.8 (Max 389.3; Min 19.9)

19.9 (Max 42.5; Min 6.4)

120.8 (Max 228.5; Min 53.8)

Loamy Sand 91 22 129.0 (Max- 415.9, Min-6.5

30.9 (Max- 61.6, Min-6.4)

132.6 (Max- 288.9; Min-67.2)

Sandy Loam 74 20 140.6 (Max 419.2; Min 13.3)

21.4 (Max 35.9; Min 6.4)

140.9 (Max 288.9; Min 65.9)

Loam 19 9 155.3 (Max 519.9; Min 6.7

35.6 (Max 53.9; Min 24.3)

140.4 (Max 241.9; Min 67.2)

Silt Loam 73 19 216.6 (Max 469.2; Min 33.4)

30.7 (Max 61.6; Min 13.5)

169.9 (Max 295.7; Min 67.2)

Clay Loam 6 0 - - -

Silty Clay Loam 18 0 - - -

Silt 2 0 - - -

Silty Clay 19 0 - - -

Total 346 81 - - -


1 Possession of Agriculture land, in bigha (1 hectare= 7.5 bigha approx.)

2 Agriculture land affected by flood, in bigha

3 Agriculture land affected by sand deposition, in bigha

4a Agriculture land leased in for cultivation, in bigha

4b Terms and conditions of leasing in

4c Quality of leased in land- sandy-1, flood prone-2,both-3

4d Leased in, within the village-1, Outside the village-2

5a Agriculture land leased out, in bigha

5b Quality of leased in land- sandy-1, flood prone-2,both-3

6a Land purchased in recent times, in bigha

6b Purpose of purchase of land

6c Buying price of land per bigha

6d Land bought within the village- 1, Outside the village-2

6e Quality of bought land- sandy-1, flood prone-2,both-3

7a Agriculture land sold, in bigha

7b Sell price of land per bigha

7c Quality of land sold- sandy-1, flood prone-2,both-3

8a Mortgage in land, if yes, quantity in bigha and price paid

8b Mortgage out land, if yes quantity in bigha and price got

9a Total land under paddy- own and leased in, in bigha

9b Total production of paddy

10 Agriculture land converted for other use/mention types

11a HH members migrated out for job, if yes, where

11b Type of job presently engaged

Annexure-I: Household Survey Schedule for Listing (stage- I)

SANDEE sponsored study on

Valuing Damages of Flood Induced Sand Deposition in Dhemaji, India

Village: Block:

Head of the household:

25


Annexure-II: Household Schedule for Detailed Survey SANDEE sponsored study on

Valuing Damages of Flood Induced Sand Deposition in Dhemaji, India

Information on the VillageName of the Respondent: 2. Village: 3. Block

A. Information on the Household

S. No

Relation with head of the HH

Sex Age Educational Attainment level#

Occupation and Income

Primary*(engaged during larger part of the year)

Place of work$

Commute@ Income per month^^

Secondary* Place of work $

Income per month ^^

1 Head

2

3

4

5

6

# Illiterate- 1, literate- 2, completed primary-3, completed middle school-4, completed high school-5, completed higher secondary-6, technical and profession diploma- 7 (specify), general graduate-8, technical graduate-9, medical graduate- 10, law graduate- 11, management graduate- 12, post-graduate and above-13, others (specify) -99

* Cultivation - 1, agriculture labour-2, livestock rearing-3, poultry- 4, fishing- 5, forestry/logging- 6, mining (sand etc)-7, manufacturing agriculture based-8, manufacturing non-agriculture based-9, construction work, road/buildings- 10, trade/shop/restaurants etc.- 11, transport, communication etc-12, finance, insurances and business services -13, community and social services-(clerical and managerial jobs in administration, jobs in education and health sector; community development, security services)-14 personal services (laundry, tailoring, hair cutting etc. specify)-15, student- 16, unemployed- 17, too old/too young to work- 18, rentiers, pensioners, remittances etc- 19, others- specify -99. Note- specify each type of job along with the code

$ Within the village-1, outside the village but within the block -2, outside the block but within the district-3, outside the district but within the state-4, outside the state- 5

Note- capture and name of the place of work outside the block@ Commute from home-1, stay outside- 2^^ Calculate gross yearly income for agriculture; other incomes are in month


B. L

and,

Pro

duct

ion

and

Retu

rn fr

om A

gric

ultu

re C

omm

oditi

es

Land

hol

ding

and

exp

endi

ture

in a

gric

ultu

re o

pera

tions

:

Ow

ned

land

(in

bigh

a, 3

big

ha =

one

acr

e)__

____

_; L

ease

d ou

t lan

d___

____

__; M

ortg

aged

land

___

____

__; L

and

sold

in la

st th

ree

year

s __

____

___;

Lea

sed

in la

nd

____

___

Type

of

poss

essi

on

(exc

lude

le

ased

out

and

m

ortg

aged

pl

ots)

Area

in

Bigh

a C

rop

culti

vate

d-se

ason

1-

sum

mer

Crop

cu

ltiva

ted

Seas

on

2-w

inte

r

Net

Irr

igat

edAr

ea in

bi

gha

Tilli

ng

met

hods

: pl

ough

-1/

trac

tor-2

, bo

th-3

Dies

el

used

in

tillin

g an

d irr

igat

ion-

in li

ters

Fert

ilise

r use

d in

kg

in

each

sea

son

Hire

d la

bour

us

ed in

nu

mbe

rs

Cost

on

la

bour

Appr

ox S

and

Dept

h,

perc

eive

d by

th

e fa

rmer

(in

inch

es)

Prod

uctio

n in

Qui

ntal

Pr

ocur

emen

t of p

addy

fr

om m

arke

t in

case

hom

e pr

oduc

tions

do

not s

uffice

(in

quin

tal)

Seas

on 1

Seas

on 2

Seas

on 1

Seas

on 2

Ow

n La

nd-

Plot

s

Hom

est

ead

Plot

1

Plot

2

Plot

3

Plot

4

Tota

l

Leas

ed

in

Plot

1

Plot

2

Plot

3

Tota

l

27


B. L

and,

Pro

duct

ion

and

Retu

rn fr

om A

gric

ultu

re C

omm

oditi

es

Land

hol

ding

and

exp

endi

ture

in a

gric

ultu

re o

pera

tions

:

Ow

ned

land

(in

bigh

a, 3

big

ha =

one

acr

e)__

____

_; L

ease

d ou

t lan

d___

____

__; M

ortg

aged

land

___

____

__; L

and

sold

in la

st th

ree

year

s __

____

___;

Lea

sed

in la

nd

____

___

Type

of

poss

essi

on

(exc

lude

le

ased

out

and

m

ortg

aged

pl

ots)

Area

in

Bigh

a C

rop

culti

vate

d-se

ason

1-

sum

mer

Crop

cu

ltiva

ted

Seas

on

2-w

inte

r

Net

Irr

igat

edAr

ea in

bi

gha

Tilli

ng

met

hods

: pl

ough

-1/

trac

tor-2

, bo

th-3

Dies

el

used

in

tillin

g an

d irr

igat

ion-

in li

ters

Fert

ilise

r use

d in

kg

in

each

sea

son

Hire

d la

bour

us

ed in

nu

mbe

rs

Cost

on

la

bour

Appr

ox S

and

Dept

h,

perc

eive

d by

th

e fa

rmer

(in

inch

es)

Prod

uctio

n in

Qui

ntal

Pr

ocur

emen

t of p

addy

fr

om m

arke

t in

case

hom

e pr

oduc

tions

do

not s

uffice

(in

quin

tal)

Seas

on 1

Seas

on 2

Seas

on 1

Seas

on 2

Ow

n La

nd-

Plot

s

Hom

est

ead

Plot

1

Plot

2

Plot

3

Plot

4

Tota

l

Leas

ed

in

Plot

1

Plot

2

Plot

3

Tota

l

B. L

and,

Pro

duct

ion

and

Retu

rn fr

om A

gric

ultu

re C

omm

oditi

es (c

ontin

ued)

Type

of p

osse

ssio

nCo

st o

f Irr

igat

ion-

die

sel,

pum

p se

t hi

ring

etc

in R

s.

Cost

of h

iring

pow

er

tille

r in

Rs.

Cost

on

Fert

ilize

rs

in R

s.

Cost

of H

YV S

eeds

in

Rs.

C

ost o

f oth

er s

eeds

in

Rs.

O

ther

cos

ts in

any

in R

s.

Ow

n La

nd-P

lots

H

ome

stea

d

Plot

1

Plot

2

Plot

3

Plot

4

Tota

l

Leas

ed in

Pl

ot 1

Plot

2

Plot

3

Tota

l


Mea

sure

take

n fo

r lan

d im

prov

emen

t and

exp

endi

ture

incu

rred

Type

of p

osse

ssio

nM

easu

res

for

prot

ectio

n;

emba

nkm

ent e

tc

Mea

sure

s fo

r rem

oval

of

sand

y la

yers

M

easu

res

for

irrig

atio

n U

se o

f mor

e do

ses

of fe

rtili

zers

Co

nver

sion

of l

and

for a

ltern

ativ

e us

e/

fishe

ry, n

on-fa

rm u

se,

plan

tatio

n et

c

Oth

er m

easu

res

(spe

cify

)To

tal l

abou

r in

volv

ed in

land

im

prov

emen

t m

easu

res

Tota

l exp

endi

ture

in

curr

ed

Yes-

1,N

o-2

Exp.

In

Rs.

Yes-

1,N

o-2

Exp.

In R

s.

Yes-

1,N

o-2

Exp.

In

Rs.

Yes-

1,N

o-2

Exp.

In

Rs.

Yes-

1,N

o-2

Exp.

In R

s.

Yes-

1,N

o-2

Exp.

In R

s.

0wn

Hire

d

Ow

n La

nd-

Plot

s H

ome

stea

d

Plot

1

Plot

2

Plot

3

Plot

4

Tota

l

Leas

ed in

Pl

ot 1

Plot

2

Plot

3

Tota

l

29


D.

Mob

ilisa

tion

of C

redi

t

Loan

Am

ount

of l

oan

take

n in

Rs.

Pu

rpos

e of

loan

#So

urce

W

hen

the

loan

was

ta

ken

Term

s of

the

loan

Inte

rest

rate

per

mon

th/

year

Amou

nt re

paid

Bala

nce

1 2 * So

urce

s- B

anks

, Coo

p so

ciet

y, M

oney

lend

ers,

Frie

nds

and

Rela

tives

#

Land

con

serv

atio

n- 1

, inv

estm

ent i

n ag

ricul

ture

inpu

ts-2

, cle

arin

g on

san

d de

posi

tion-

3, h

ouse

hold

con

sum

er n

eeds

-4, m

edic

al tr

eatm

ents

nee

d- 5

, edu

catio

nal n

eeds

-5, o

ther

s (s

peci

fy) 9

E. M

igra

tion

and

Rem

ittan

ces

HH

mem

bers

- mig

rate

d ou

tCo

de (f

rom

HH

ta

ble)

Reas

ons

for d

ecis

ion

on m

igra

tion

(rec

ord

mul

tiple

resp

onse

s) *

Ti

me

sinc

e ou

t mig

rate

d Pl

ace

of w

ork

–spe

cify

M

onth

ly e

arni

ng in

Rs.

Re

mitt

ance

in a

mon

th, i

n Rs

.

1 2 3 * Re

turn

from

agr

icul

ture

doe

s no

t suffi

ce -1

, low

/ins

uffici

ent r

etur

ns fr

om n

on-a

gric

ultu

re jo

bs a

vaila

ble

loca

lly-2

, lac

k of

fina

ncia

l res

ourc

es to

inve

st in

sel

f-em

ploy

men

t ini

tiativ

es-3

, int

eres

ted

in s

ervi

ce s

ecto

r job

s - 4

, in

tere

sted

in jo

bs in

urb

an a

reas

- 5,

othe

rs (s

peci

fy)-

9


F. H

H a

vaili

ng b

enefi

ts u

nder

flag

ship

pro

gram

mes

of g

over

nmen

t

Sche

mes

Spec

ify th

e ty

pes

of b

enefi

ts

Estim

ated

mon

etar

y ga

in/r

etur

n fr

om th

e co

ntin

uing

act

iviti

es in

dur

ing

the

last

yea

r; fo

r one

tim

e gr

ant r

ecor

d th

e on

e tim

e be

nefit

s re

ceiv

ed

Empl

oym

ent g

uara

ntee

sch

eme-

NRE

GA

Hou

sing

- IAY

Tota

l San

itatio

n Ca

mpa

ign

Hea

lth- N

RHM

Child

dev

elop

men

t Ser

vice

s- IC

DS

Sarv

a Si

ksha

- Uni

vers

al E

duca

tiona

l Pro

gram

me

Mid

Day

Mea

l Sch

eme

for S

choo

l Chi

ldre

n

Soci

al A

ssis

tanc

e Pr

ogra

mm

e- O

ld a

ge p

ensi

ons,

sub

sidi

zed

ratio

ns e

tc.

Rura

l ele

ctric

ity

Floo

d da

mag

e re

late

d as

sist

ance

Oth

ers

(spe

cify

)

31


SANDEEP.O. Box 8975, E.P.C 1056, Lalitpur, Nepal

Street address: c/o ICIMOD, Khumaltar, Lalitpur, Nepal

Tel: 977 1 5003222, Fax: 977 1 5003299, Email: [email protected] Web: www.sandeeonline.org

SANDEE Sponsors

Sand Deposition and Poverty among Farm Households

Documents

Transcript of Sand Deposition and Poverty among Farm Households